Read on Twitter
Impact of #COVID19 measures on electricity consumption, published in @iScience_CP @CellPressNews
Cumulative impact March-July ~3-11% reduction
Interactive results & data: https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
Daily results per country
1/12
Cumulative impact March-July ~3-11% reduction
Interactive results & data: https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
Daily results per country
1/12
Daily impact of COVID-19 measures on electricity consumption in Austria
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
Cum. impact: -4.6%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
2/12
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
Cum. impact: -4.6%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
2/12
Daily impact of COVID-19 measures on electricity consumption in California
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
Cum. impact: -4.7%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
3/12
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
Cum. impact: -4.7%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
3/12
Daily impact of COVID-19 measures on electricity consumption in Florida
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
No net negative impact (+0.9%)
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
4/12
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
No net negative impact (+0.9%)
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
4/12
Daily impact of COVID-19 measures on electricity consumption in France
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
Cum. impact: -8.8%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
5/12
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
Cum. impact: -8.8%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
5/12
Daily impact of COVID-19 measures on electricity consumption in Germany
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
Cum. impact: -7.5%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
6/12
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
Cum. impact: -7.5%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
6/12
Daily impact of COVID-19 measures on electricity consumption in Great Britain
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
Cum. impact: -11.4%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
7/12
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
Cum. impact: -11.4%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
7/12
Daily impact of COVID-19 measures on electricity consumption in Italy
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
Cum. impact: -11%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
8/12
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
Cum. impact: -11%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
8/12
Daily impact of COVID-19 measures on electricity consumption in New York
Left: baseline (black) vs actual consumption (color)
Right: daily impact
Cum. impact: -3.3%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
9/12
Left: baseline (black) vs actual consumption (color)
Right: daily impact
Cum. impact: -3.3%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
9/12
Daily impact of COVID-19 measures on electricity consumption in Spain
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
Cum. impact: -8.6%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
10/12
Left: baseline (black) vs actual consumption (color)
Right: daily % impact
Cum. impact: -8.6%
Interactive figures:
https://jlprol.shinyapps.io/covid/
Paper: https://www.sciencedirect.com/science/article/pii/S2589004220308312
10/12
Relationship consumption decline-stringency of the COVID-19 measures.
Stringency index by @annajessiep, @beakira, @angrist_noam, @thomasnhale, @TobyMPhillips & team at @BlavatnikSchool @UniofOxford
https://www.bsg.ox.ac.uk/research/research-projects/coronavirus-government-response-tracker
11/12
Stringency index by @annajessiep, @beakira, @angrist_noam, @thomasnhale, @TobyMPhillips & team at @BlavatnikSchool @UniofOxford
https://www.bsg.ox.ac.uk/research/research-projects/coronavirus-government-response-tracker
11/12
The method to estimate the baseline is a dynamic harmonic regression.
Seasonality modelled with Fourier terms, ARIMA errors for short-term dynamics, Box-Cox trans. for variance stabilisation, holidays dummy and quadratic temperature.
Test set mean abs. % error 2.7–4.6%
12/12
Seasonality modelled with Fourier terms, ARIMA errors for short-term dynamics, Box-Cox trans. for variance stabilisation, holidays dummy and quadratic temperature.
Test set mean abs. % error 2.7–4.6%
12/12
PS: if this is useful, we will update the results and extend the analysis to other countries
